Comment by kortilla
1 day ago
It’s not, or at least it definitely wasn’t a year ago. Those cars were something like $700k each and then there is a lot of software dev and AI infra to pay for. They were charging more than Lyft and were still losing money per ride.
> Those cars were something like $700k each
Source? Also, those numbers are quoted with including capital expenditures, which will get depreciated over years. Also, to do a like for like comparison you have to include all of the people, processes, and systems in place that support drivers. That cost center at Uber has to be in the millions globally for Uber (and hard to calculate on a per driver basis, but possible if we had internal numbers).
No public source, just talking to old coworkers at Google asking why the rollout was so incredibly slow in the Phoenix location. It’s almost all one big county and the driving conditions are the same across that whole valley.
So why not open up the entire service area? The answer was cost. At the time they were still losing money per ride to help get ML training feedback.